IA & Pedagogy (EN)

IA & Pedagogy

An interactive guide

Let's go!

Introduction

This interactive guide is designed to provide you with ideas and recommendations for more effectively integrating the challenges of generative AI into your teaching practices. It is organized around three main themes: AI and Assessment, AI and Best Practices, and AI as a Teaching Assistant. Click the button below to access the menu!

Start

The K-lab training offer

Index

AI as a TA

AI & best practices

AI & assessment

This section aims to provide guidance on best practices that faculty members can share with students to support responsible and effective use of AI tools.

This section is designed to offer you ideas for rethinking your assessment strategies with AI in mind.

This section is designed to show you how to use AI as a "teaching assistant" in your daily practice.

Section 1. AI & assessment

At first, generative AI acted as a catalyst for existing challenges, particularly issues like cheating during online exams and a lack of personal engagement. However, generative AI is now increasingly seen as an opportunity to rethink and improve teaching, learning, and, of course, assessment practices. In this section, we will share ideas for adapting your assessments to the new challenges posed by AI.

Section 1 - Index

AI & assessment

Some precautions to take for mandatory summative assessments.

Implement formative and continuous assessments, and incorporate AI into course syllabi.

Place greater emphasis on skills development, particularly soft skills.

Info

Info

Info

ESSEC AI Study Group's guidelines

Rethinking assessments and syllabi

These different activities refocus learners' attention on the content and the learning process, rather than solely on the outcome (grades), significantly reducing attempts at plagiarism and the automatic generation of answers. In addition, implementing continuous assessments is another promising approach — for example, using a flipped classroom model, where students review resources at home beforehand and then participate in in-class workshops focused on reflection and practice, with as much supervised group work as possible. This system offers a double benefit: it helps reduce cheating and the misuse of AI while also maintaining students' attention and motivation. Finally, it may also be valuable to incorporate AI into course syllabi by designing exercises such as generating a first draft with ChatGPT, followed by a guided self-assessment and peer commentary activity in class. To support you, ready-to-use Learning Journeys are available for download on the FeedbackFruits website.

One initial approach to effectively counter the inappropriate use of generative AI is to rethink your assessment strategy. Several experiences, particularly with SPOCs, have shown the effectiveness of formative assessments: formative quizzes with detailed feedback for each correct or incorrect answer, self-assessment, and peer assessment prior to the submission of a final assignment.

Emphasize soft skills

Implementing competency-based assessment involves defining key skills — particularly soft skills such as communication, collaboration, critical thinking, and adaptability — and designing activities that allow these skills to be observed in action, such as collaborative projects or case studies, ideally carried out during class time with the teacher acting as a mentor or coach. By using evaluation rubrics with specific criteria and mastery levels, you can measure each student's progress in a detailed and structured way. Involving learners through self-assessment and peer or intra-group evaluation, including writing and incorporating constructive feedback, also helps students become more aware of their strengths and areas for improvement — an important part of developing and valuing the soft skills themselves.

Click the image to enlarge it!

Tools implemented at ESSEC, such as FeedbackFruits (with activities like peer assessment, intra-group evaluation, and self-assessment), as well as updated theoretical models adapted for the AI era (see image above), can help you design appropriate activities.

Exam precautions

If you wish to set up an online exam, you can use a screen blocker (Safe Exam Browser) on Moodle (note that this solution requires prior setup and familiarization). If your exam is open-book, you may allow each student to prepare a personal notes sheet ("cheat sheet"), which enables the use of Safe Exam Browser while still granting access to key reference materials. Another option is the "Offline Quiz" feature, which allows you to create multiple-choice quizzes (MCQs) on Moodle, print them, and then automatically correct them by scanning the completed answer sheets. Regarding exam content, if your subject matter allows, it can be valuable to ask students to connect the topics you wish to assess with a personal or professional experience — past or future — encouraging reflection and projection, and at the same time developing soft skills. Finally, it is recommended to assess students using recent topics that are not yet widely covered by generative AI tools.

For mandatory summative assessments (such as mid-term and final exams), it is preferable — when the context allows (in terms of student numbers and exam length) — to favor paper-based exams. This helps prevent both traditional forms of cheating and the use of AI tools. It also gives students the opportunity to practice handwriting, which supports long-term memory retention and knowledge acquisition. Whenever possible, encourage oral examinations, including participation, presentations, and defenses.

Section 2. AI & Best practices

This section aims to equip you with practical guidelines and suggestions to share with your students regarding their use of Gen AI. Based on the ESSEC position paper, it presents a comprehensive framework with targeted recommendations to promote responsible, ethical, and informed use. Additionally, it offers a clear methodology to assist you in identifying potential misuse of AI in student assignments.

Section 2 - Index

AI & Best practices

Identifying fraudulent use of Generative AI

ESSEC position paper

Key points to relay to students

Info

Info

Info

ESSEC position paper

• AI offers opportunities to enhance pedagogical approaches.
• The impact of AI on organizations and society is a key research focus at ESSEC.
• AI serves as a lever to optimize processes and operational efficiency.
• Risk management and the preservation of academic integrity rely on implementing tools and sharing best practices.
• Successful adoption of generative AI at ESSEC depends on a comprehensive training program and support for all members of the community (students, participants, faculty, and administrative staff).

• ChatGPT and its competitors rely on a model designed to generate plausible—but not necessarily accurate—results. Avoid using generative AI in contexts where accuracy is critical unless you have the expertise to evaluate the outputs.
• Lack of transparency regarding the tool’s knowledge base (dataset composition): What content? What coverage?
• Hallucinations / Bias / Copyright: the tool’s limitations. The probabilistic model can lead generative AI to produce outputs that may reproduce stereotypes or “dominant” viewpoints, often relying on untraceable or uncredited content (risk of plagiarism).
• Use in information retrieval: generative AI is not a search engine. Search engines provide a curated selection of resources to answer a query, whereas generative AI produces text based on a probabilistic and assertive model.

What framework and guidelines should be given to students? (1/2)

Prerequisites for a controlled and reasoned use of generative AI

Generative AI produces content (text, images, video, etc.) based on a largely probabilistic and assertive model.

What framework and guidelines should be given to students? (2/2)

Use these tools for what they are: examples of appropriate applications

Document research: Identify keywords and key concepts relevant to the topic at hand. Writing assistance: Occasional rephrasing to strengthen an argument or improve the flow of a text. Ideation/Brainstorming: Formulate hypotheses, explore research ideas, or obtain suggestions on topics to stimulate creativity. Summarizing a copyrighted document (article, academic paper, study, etc.) using generative AI constitutes academic misconduct. The analyzed content is integrated into the service’s knowledge base without prior authorization.

Promoting Integrity and Transparency

• It is strongly recommended that every instructor clearly specify, within their guidelines, the conditions and modalities for the use of generative AI tools.
• Any student or participant who uses AI in the context of their academic work (theses, presentations, etc.) must explicitly disclose this in order to ensure transparency and uphold academic integrity. This requirement applies in all cases, whether the tool was used as a source (cited in bibliographic references and/or within the text) or as a support tool (writing assistance, brainstorming, etc.).

Info on Magister +

Detecting Fraudulent Use of AI: The Importance of a Constellation of Evidence

AI detection tools such as Magister+ by Compilatio should not be the only solutions to be considered. These tools flag potential suspicions of fraud, indicating a probability, but never provide objective proof. The prevailing approach relies on a constellation of evidence, as no technical solution is perfect. The feedback from the reader/reviewer is the primary signal to consider. Recommended procedure:

• Suspicion raised by the reader/reviewer;
• Verification of the problematic report using Magister+ (Compilatio);
• Discussion with the student.

Points of Vigilance for the Instructor (Analysis Grid)

Form: style and structure

Content

Quality of the bibliography / references

Section 3. AI as a TA

The goal of this section is to show you several ways to use generative AI as an ally in your teaching practices — from syllabus design to grading student work, as well as creating learning activities, evaluation rubrics, or ongoing case studies. When used appropriately, AI can save you considerable time, much like a teaching assistant would.

Section 3 - Index

AI as a TA

Support for automated grading and report generation

Support for creating teaching resources and learning activities

Enabling AI Assistance features on our various tools

Info

Info

Info

For example, you can provide your chosen AI tool with a first draft of your syllabus and ask it to rewrite the learning outcomes, organizing them into three progressive levels based on Bloom’s taxonomy. You can also share a general outline of your planned sessions and ask the AI to suggest assessment activities aligned with the previously defined objectives, and/or propose ideas for ongoing case studies. Additionally, you can use generative AI to design formative quizzes, with detailed feedback for each answer and general feedback for each question and for the quiz as a whole. You would then review and refine the AI’s draft to correct and adjust its suggestions. Generative AI is also effective for creating and improving assessment grids, generating detailed criteria with several levels of mastery and clear descriptions of expected performance at each level, according to your instructions. Don’t hesitate to contact the K-lab's instructional designers if you would like support with this!

Support for creating educational resources and learning activities

Generative AI can be seen as a true teaching assistant, supporting you in the design and improvement of your courses — from the syllabus to the final assessment, including all intermediate activities. When used appropriately, it can save you a considerable amount of time!

Feedback from the AI Feedback Coach appears next to the comment field as students write their evaluations, offering real-time suggestions to improve their work based on the evaluation criteria, the language used, and the type of evaluation. Suggestions address both form (such as language, length, etc.) and content (such as providing constructive comments). This feature encourages students to engage in an initial self-correction process before you review their work yourself, providing them with an opportunity to strengthen their writing skills and develop key soft skills: adopting a reflective attitude toward their own work and work processes (whether individual or group-based, depending on the assignment), and fostering an empathetic and supportive approach toward themselves and others.

Enabling the "AI-Generated Feedback Coach" for online assignments

The FeedbackFruits tool, integrated into Moodle through three activities (peer assessment, intra-group evaluation, and self-assessment), offers an AI-generated Feedback Coach option in its settings. This feature supports students as they complete their evaluations and helps them improve their work.

More info here: https://feedbackfruits.com/automated-feedback

Generative AI can also help you summarize and/or analyze student work, using prompts such as: "Identify the main arguments in this assignment and evaluate their relevance [based on specific criteria, etc.]." In addition, AI can generate complete documents that you can download, such as reports, based on the data you provide (for example, extracted activity reports from Moodle) and clear prompts like: "Draft a [specific format, length] report on the class’s performance this semester [organized into a certain number of thematic categories, etc.]." However, it’s important to keep in mind that generative AI has limitations: it may lack nuance, and its suggestions should be reviewed to ensure their accuracy and pedagogical relevance. From an ethical standpoint, you should also consider the potential circulation of any data submitted for analysis. In short, AI offers a useful first draft or complementary tool, but it does not replace the teacher’s human expertise.

Support for automated grading of assignments and report generation

A true ally and virtual teaching assistant, generative AI can help you produce a first draft of structured feedback for your students — for example, by using a prompt such as: "Provide constructive feedback for this essay: [attachment], highlighting strengths and areas for improvement [etc.]."

The K-lab training offer

Partnering with ChatGPT! A guide for educators

Creating Your First Visuals with AI

Generative AI in Practice: Getting Started

Audience: Students/Staff

Audience: Students/Staff

Audience: Teachers

Objectives / Content

• Fundamentals of the main generative AI tools
• Best practices for safe and responsible use
• Introduction to prompting basics

Objectives / Content

• Introduction to key free AI tools
• Guidelines for safe and responsible use
• Essentials of effective prompting
• Copyright issues and ethical considerations

Objectifves / Content Coming soon

Registration

Coming soon

Registration

Recourse to Magister+ In cases of doubt or suspicion regarding the authenticity of a student’s work, instructors may use Magister+ as a support tool. The K-lab has subscribed to the “AI Detection” option from Compilatio (the anti-plagiarism solution deployed at ESSEC), thereby upgrading from the Magister software (similarity detection) to Magister+ (detection of similarities and AI-generated content). Service Overview For each analysis, Magister+ provides an overall percentage of suspect content, including a similarity score (plagiarism), a score for unrecognized languages, and a score for potentially AI-generated content. The latter, labeled “AI-generated text,” indicates a suspicion—a likelihood of use—but does not constitute objective proof. Usage Guidelines The software flags potentially suspect text segments but does not provide visual proof. It is up to the instructor to confirm or dismiss the suspicion. A discussion with the student is essential in such cases.

For any questions regarding the use of the tool, please contact us: compilatio@essec.edu

Tutorial :

Points of Vigilance

Content

• Lack of depth and critical reflection: the treatment of the subject is superficial and generic.
• Sense of padding and verbosity: excessive descriptive elements with little analysis.
• Mismatch between the content and the target audience.

Points of Vigilance

Form, style, and structure

• Excessive neutrality in the text
• Abrupt shifts in style
• Repetitions or inconsistencies in the use of certain terms
• Overuse of formal transition phrases
• Absence of spelling mistakes

Points of Vigilance

Quality of the bibliography

• Incorrect bibliographic references (fabrications, inaccuracies)
• Bibliographic references of excessive quality or not aligned with the student’s academic background
• Bibliographic references inaccessible via the documentary resources available